How to control smoothing

[EmilioKolo]

Hi,

I'm using joypy to display the position of mutations associated with different diseases over a gene's length, and I need some tight control over the smoothing parameter of the ridgeline plot.

Example table with dummy data:

mutation	position	disease
p.Cys100Arg	100	D1
p.Gly120Arg	120	D1
p.Trp122Ser	122	D1
p.His60Gln	60	D2
p.Gln110His	110	D2
p.Arg65Cys	65	D3
p.Arg67Gln	67	D3
p.Ser70Trp	70	D3

When I run the joyplot function, I set up x_range to be the length of the entire gene (+/- ~10 to fine-tune the x axis). So, if the gene is 150 aminoacids long, I set x_range to [-10, 160]. E.g.

joypy.joyplot(dataframe_table, by='disease', column='position', x_range=(-10,160))

This leads to a way too smooth graph where far-away mutations appear to be "connected" by their frequencies, so D2 looks like a smooth cap over the entire graph, when it's only two mutations at positions 60 and 110.

I'm assuming that the default functionality searches for values around each point in the x axis and makes some transformation to smooth things out. I would like to have some control over that smoothing factor, since the influencing of a value in the x axis by a different value that's too far away is counterproductive for my purposes.

The data points can be quite sparse in many regions and for the most uncommon diseases so a histogram would not be ideal, since bins are fixed boxes with no relevance to the function of the gene, while every point being slightly influenced by nearby mutations has some biological relevance... up to a point.

I tried to use the kind="values" workaround but that would require some reshaping of the data and it may not look as good, but if this isn't a planned feature I will probably just do that.

Thanks in advance.